plannotator-review

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). The SKILL.md instructs running "plannotator review [optional-pr-url]" to open a browser-based code review for a (potentially arbitrary) pull request URL and to "act on the feedback" returned, meaning the agent will fetch and interpret untrusted, user-generated PR/review content as part of its workflow.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill runs the external command "plannotator review [optional-pr-url]" at runtime and will fetch and act on the content of a user-supplied pull request URL (e.g. https://github.com/org/repo/pull/123), so remote PR content can directly control the agent's prompts and behavior.